In my first article, Celsion's HEAT Study: A Far From Certain Outcome, I delineated why one should be cautious in interpreting Celsion's (NASDAQ:CLSN) ThermoDox Phase I trial as a proxy for a potential Phase III success. Even the most bullish of investors would ultimately concede that extrapolating from small datasets to larger trials presents a challenge. Nonetheless, despite the absence of robust ThermoDox data, arguments have continued to highlight literature that supports a potential role for doxorubicin efficacy in the treatment of unresectable hepatocellular carcinoma (HCC) in patients with greater than 3 cm tumors. Those include the use of chemotherapy as part of Transcatheter Arterial Chemoembolization (TACE) procedures as well as doxorubicin-eluting bead (DEB) embolization. More recently, a published meta-analysis suggesting a survival benefit for radiofrequency ablation (RFA) plus TACE over RFA alone has been highlighted as evidence of potential ThermoDox Phase III trial success.
In this article, I will highlight arguments that demonstrate a marginal role, if any, for doxorubicin in the treatment of HCC. Further, I will refute the parallels between TACE, DEB-TACE and ThermoDox, adding to the body of evidence that the HEAT trial will most likely fail.
The following are my specific observations:
1. Benefits of chemotherapy in TACE are unclear. TACE derives its effect by interrupting arterial blood supply to tumors and potentially by focused administration of chemotherapy. Notably, the actual benefit of chemotherapy in HCC is still debated. Specifically, a meta-analysis of TACE versus bland embolization (without using any chemotherapeutic agents) demonstrated no survival differences between the two techniques.
2. Benefit of regional sustained doxorubicin is marginal. Prospective studies of doxorubicin-eluting bead embolization exhibit non-statistically significant superior trends to traditional TACE. This potential marginal effect is achieved through a 14-day sustained doxorubicin release.
3. ThermoDox mechanism of action and pharmacokinetics profile points to limited doxorubicin exposure and hence unlikely efficacy. I estimate that effective intratumoral doxorubicin concentrations achieved with ThermoDox + RFA are at a maximum about 9 times less than with the DEB-TACE technique. Further, Celsion's drug exhibits peak doxorubicin concentrations at less than 1 hour post administration and almost complete elimination by 6 hours. Taking into account that much higher sustained regional chemotherapeutic exposure over days has only a marginal effect, it is doubtful that short exposure to ThermoDox will demonstrate any benefit in ongoing HCC trials.
Benefits of chemotherapy in TACE are unclear
Transcatheter arterial embolization (TAE), known as bland embolization, is a minimally invasive image-guided therapy for treating intermediate-stage HCC. It is performed via a percutaneous transarterial approach. A catheter is passed through the common hepatic artery, into the proper hepatic artery to reach the target branches that are supplying blood to the tumor. When TAE is combined with prior injection into the hepatic artery of a chemotherapeutic agent (being most often doxorubicin at doses of 50-75 mg/m2), the procedure becomes known as Transcatheter Arterial Chemoembolization (TACE).
The benefits of TACE are viewed theoretically as twofold: Arterial embolization interrupts blood supply to the tumor, thus causing tumor cell death and suppressing its growth. Secondly, local administration of chemotherapy allows for higher doses than systemic exposure would normally permit. Further, drug presence is extended in the tumor bed due to decreased hepatic perfusion subsequent to embolization.
Notably, TAE, in the absence of a chemotherapy agent, is known to result in objective treatment responses, highlighting the important role of embolization in depriving nutrient and oxygen supply to tumor cells, without worsening the patient's liver function. The question that arises from this observation is: What is the incremental benefit of adding chemotherapy to the TAE technique? Interestingly, the actual contribution of chemotherapy to TAE has been explored in the literature by a meta-analysis performed by Marelli et al., (2007), evaluating 3 randomized, controlled studies with a total of 412 patients.
Strikingly, the study's authors concluded that there is no evidence of an additive or synergistic effect with the addition of chemotherapy to TAE. Specifically, chemoembolization versus embolization alone demonstrated no survival difference. Based on the evidence, they go on to conclude that ischemia resulting from embolization is likely the main factor inducing reduction in tumor size in TACE; an observation which is consistent with the knowledge that HCC is notoriously insensitive to chemotherapeutic drugs.
Benefit of regional sustained doxorubicin is marginal
In recent years doxorubicin-eluting-bead (DEB) technology (DC Bead) has been developed to enhance local tumor drug delivery while reducing systemic availability. DEB technology allows for controlled doxorubicin release (150 mg per procedure) over a period of up to 14 days with peak concentrations achieved by day 3.
The use of DEB has been explored in a number of studies with the most notable being PRECISION V (Prospective Randomized Study of Doxorubicin in the Treatment of Hepatocellular Carcinoma by Drug-Eluting Bead Embolization).
This international, multicenter, prospective, randomized, single-blind 212 patient trial compared conventional TACE to DEB-TACE with a primary endpoint of tumor response at 6 months. The DEB-TACE arm showed marginal trends of higher rates of complete response, objective response, and disease control compared with standard TACE (27% vs. 22%, 52% vs. 44% and 63% vs. 52%, respectively). Those, however, were not statistically significant to prove actual superiority (one sided P=0.11). Improved tolerability and reductions in liver toxicity were observed and shown to be statistically significant (p<0.001).
Notably, in this study, the traditional TACE group received an intra-arterial injection of an emulsion of doxorubicin of 90-150 mg. This compares to sustained delivery of 150 mg doxorubicin by the use of DEB. While the benefit of DEB was marginal, the conclusion drawn from the above study suggests that sustained delivery of similar doxorubicin doses over 14 days might offer a slight benefit over hepatic bolus chemotherapy with superior tolerability.
ThermoDox mechanism of action and pharmacokinetics profile points to limited doxorubicin exposure
ThermoDox is a heat sensitive liposomal formulation of doxorubicin. It selectively releases its contents when exposed to temperatures ≥ 39.5 ° C. Liposomes are rapidly concentrated and cleared by the spleen and liver, hence, preferentially accumulating in these organs.
Pharmacokinetic studies of ThermoDox reveal the following: (Figure 1)
- Maximum plasma concentration occurs just before the end of a 30-min infusion.
- Liposomal doxorubicin represents 56% of total plasma doxorubicin.
- The major portion of exposure to the drug occurs during the first 6 hours after infusion.
Figure 1. ThermoDox Pharmacokinetics
The HEAT trial explores the potential benefit of 50 mg/m2 ThermoDox infusion in addition to RFA. According to the trial protocol, RFA will be initiated approximately at a minimum of 15 minutes after the initiation of study drug infusion and should be completed no later than 3 hours after study drug infusion initiation.
Based on the drug's pharmacokinetic profile, the following observations can be made:
- With ThermoDox dosing at 50mg/m2 and a mean body surface area of about 1.75 m2, patients on average are dosed with 87.5 mg of study drug.
- Patients are expected on average to be exposed to 38.5 mg of systemic free doxorubicin and 49 mg liposomal doxorubicin (being 56% of 87.5 mg ThermoDox).
- Release of doxorubicin in the liver would be achieved by liposomes exposed to high temperatures due to the RFA procedure. i.e. Only a fraction of doxorubicin from the maximum of 49 mg of liposomal doxorubicin accumulating in the liver is actually released in the vicinity of treated tumors. The remainder of the chemotherapy drug is inert, due to its liposomal encapsulation and is ultimately cleared from the body (see Figure 2).
Figure 2. Temperature-dependent doxorubicin release from ThermoDox
Source: Company presentation
It is therefore evident that the effective chemotherapy concentration delivered within the liver (or to the tumor area) is but a small fraction of the potential 49 mg of liposomal doxorubicin with which patients are dosed. The following are my assumptions:
- The size of an average liver is 300 cm2 (20 X 15 cm).
- The maximum liver heated tumor area would be 100 cm2 (10 X 10 cm), consistent with the trial's inclusion criteria anticipating that ablation volume will be no larger than either removal of 3 hepatic segments or removal of more than 30% of total liver volume (as per maximum surgical limit).
- Using the above assumptions, maximum potential doxorubicin dose delivered regionally would be (100/300 X 49 mg) = 16.3 mg.
This ThermoDox dose contrasts with intratumoral delivery of 150 mg of doxorubicin achieved by techniques such as DEB-TACE. i.e. effective maximum regional doxorubicin concentrations delivered through ThermoDox + RFA would amount to 9 times less chemotherapy than expected with traditional techniques.
Furthermore, patients in Celsion's HEAT trial do not undergo embolization, and therefore doxorubicin clearance from their hepatic tissue is expected to be rapid. Embolization by traditional techniques decreases hepatic perfusion, hence extending drug presence in the tumor bed. Doxorubicin-eluting beads have been specifically engineered for such controlled release. By contrast, patients dosed with ThermoDox are therefore likely to be exposed to a short burst of low dose regional doxorubicin rather than sustained high concentrations of chemotherapy as used by traditional techniques.
As to free systemic doxorubicin, as mentioned above, 50 mg/m2 ThermoDox would result in 38.5 mg of free drug. This would amount to 2.7 times less exposure than with intravenous doxorubicin as used clinically with limited results.
With such transiently low intratumoral and systemic doxorubicin exposure it seems doubtful that meaningful antiproliferative responses could be seen either regionally or distally. Consequently, it is unlikely that patients would derive a clinical benefit from ThermoDox.
Addressing some bullish ThermoDox arguments:
In the context of the evidence presented above, it is worth revisiting some of the bullish arguments pertaining to the potential success of the HEAT trial:
1. Doxorubicin is a well-characterized chemotherapy agent with known efficacy in HCC
There has never been an adequately powered study of doxorubicin against best supportive care to support the routine use of this relatively toxic drug. But the largest multinational study of 60 mg/m2 doxorubicin compared with the antifolate drug nolatrexed revealed a very disappointing 2% overall response rate and a median survival time of approximately 6 months, slightly better than that seen with the experimental agent nolatrexed. Based on this result, the consensus opinion has been that it is not necessary to use doxorubicin as the control arm in a study of a new systemic agent.
Further, as discussed above, there is no clear evidence, to date, of an additive or synergistic effect by the addition of doxorubicin to TAE.
2. ThermoDox liposomal formulation allows for greater intravenous dosing than free doxorubicin
That, in fact, is blatantly untrue. In a Phase I trial among 24 subjects who received RFA + ThermoDox for HCC or metastatic liver cancer, the maximum tolerated dose (MTD) was determined to be 50 mg/m2 , based on two dose-limiting toxicities (grade 3 alanine aminotransferase increase, grade 4 neutropenia) at a dose of 60 mg/m2.
In clinical studies, and in practice, intravenous doxorubicin is actually administered at comparable doses to ThermoDox at 60 mg/m2 intravenously every 21 days.
As mentioned above, 50 mg/m2 ThermoDox would result on average in 38.5 mg of free doxorubicin. This compares to an average of 105 mg free drug when doxorubicin is dosed intravenously. i.e. ThermoDox results in 2.7 times less systemic drug exposure than in normal clinical practice with which marginal efficacy is seen.
3. ThermoDox selectively accumulates in the liver allowing greater localized doxorubicin exposure
ThermoDox's liposomal formulation of doxorubicin does support the notion of selective liver accumulation. However, as argued above, effective regional doxorubicin doses achieved in the HEAT trial are likely to be at a maximum 9 times less than expected with techniques such as DEB-TACE. Further, in the absence of embolization, clearance from the hepatic tissue is expected to be much more rapid than by traditional regional delivery methods.
4. RFA+TACE shows benefit versus RFA alone in patients comparable to those enrolled in the HEAT trial
Some have suggested that TACE RFA can be used as a proxy for the addition of ThermoDox to RFA. However, there is limited evidence to conclude that there is an additive (or synergistic) benefit of adding TACE to RFA in patients with >3cm HCC tumors. Moreover, the contribution of doxorubicin to that potential effect is even more difficult to discern. That said, as previously demonstrated, regional doxorubicin concentrations achieved by techniques such as TACE are many fold higher than could ever be achieved by ThermoDox. And as a reminder, embolization which is believed to enhance RF ablation, is not part of the ongoing HEAT trial. Therefore, concluding success in Celsion's Phase III study based on a potential RFA + TACE benefit over RFA alone is not warranted by the weight of current evidence.
Recently, a meta-analysis was published demonstrating a survival benefit for RFA + TACE versus RFA alone in HCC. Importantly, a subgroup analysis showed that the benefit is specific to patients with >3cm tumors, while no treatment advantage was observed in smaller tumors. This publication is of high relevance to Celsion's HEAT trial, since the patient population explored is comparable to the one enrolled in the Phase III study i.e. HCC with tumors >3cm.
Upon my review of this data, I would like to highlight the following points:
i) The meta-analysis included 7 randomized-controlled trials with a total of 574 patients with HCC. Of those, 5 trials involved tumors >3cm in 500 patients.
ii) The largest, most robust study included in the analysis is a 196-patient trial by Cheung et al. published in the high-impact journal of the American Medical Association (JAMA) in April 2008 entitled "Chemoembolization Combined With Radiofrequency Ablation for Patients With Hepatocellular Carcinoma Larger Than 3 cm."
Notably, however, in May 2009, that same study was retracted by the Shandong University School of Medicine. In an official letter to JAMA, university officials concluded upon investigation that: "This study was not a well designed, randomized and controlled clinical trial despite the fact that chemoembolization and radiofrequency ablation for patients with hepatocellular carcinoma have been performed in Shandong University Qilu Hospital for many years. Therefore, conclusions drawn from this study are not valid. Because of these unscientific behaviors, we suggest that the article by Dr Cheng should be withdrawn from JAMA."
The inclusion of this study in the cited meta-analysis published in 2012 is a clear omission by the authors, reviewers and editors of the European Journal of Gastroenterology & Hepatology, calling into question the study's overall integrity and conclusions. Based on my observation, this publication itself should be retracted. Specifically, exclusion of this study would result in only 4 relatively poor studies of patients with tumors >3cm in a total of only 300 subjects.
iii) Assuming RFA + TACE does provide a benefit, the mechanism proposed by the various studies' authors primarily involves a reduction of the cooling effect of hepatic blood flow on thermal coagulation by embolization prior to RF ablation. Secondarily, the effect of chemotherapy agents on cancer cells during treatment is speculated to be enhanced by hyperthermia.
While effective new treatment options for HCC are needed, my conclusion is that the HEAT trial will most likely fail. The ThermoDox concept pursued by Celsion is a theoretically elegant method to address residual micro-metastases in combination with the RFA procedure. There is, unfortunately, an overwhelming body of evidence to suggest that the company's approach is lacking on several fronts.
Most strikingly, there is an absence of evidence that sustained high dose hepatic doxorubicin provides a meaningful additive or synergistic effect to embolization techniques. As demonstrated above, ThermoDox's pharmacokinetic profile suggests a significantly lower effective intratumoral doses of doxorubicin in combination with RFA than with established traditional techniques such as TACE. Such observations bring into question the rationale behind ThermoDox and its potential effectiveness in demonstrating a meaningful clinical benefit in the ongoing Phase III trial.